1 .TH fio 1 "December 2014" "User Manual"
3 fio \- flexible I/O tester
6 [\fIoptions\fR] [\fIjobfile\fR]...
9 is a tool that will spawn a number of threads or processes doing a
10 particular type of I/O action as specified by the user.
11 The typical use of fio is to write a job file matching the I/O load
12 one wants to simulate.
15 .BI \-\-debug \fR=\fPtype
16 Enable verbose tracing of various fio actions. May be `all' for all types
17 or individual types separated by a comma (eg \-\-debug=io,file). `help' will
18 list all available tracing options.
20 .BI \-\-output \fR=\fPfilename
21 Write output to \fIfilename\fR.
23 .BI \-\-output-format \fR=\fPformat
24 Set the reporting format to \fInormal\fR, \fIterse\fR, \fIjson\fR, or
25 \fIjson+\fR. Multiple formats can be selected, separate by a comma. \fIterse\fR
26 is a CSV based format. \fIjson+\fR is like \fIjson\fR, except it adds a full
27 dump of the latency buckets.
29 .BI \-\-runtime \fR=\fPruntime
30 Limit run time to \fIruntime\fR seconds.
33 Generate per-job bandwidth logs.
36 Print statistics in a terse, semicolon-delimited format.
39 Print statistics in selected mode AND terse, semicolon-delimited format.
40 Deprecated, use \-\-output-format instead to select multiple formats.
43 Display version information and exit.
45 .BI \-\-terse\-version \fR=\fPversion
46 Set terse version output format (Current version 3, or older version 2).
49 Display usage information and exit.
52 Perform test and validation of internal CPU clock
54 .BI \-\-crctest[\fR=\fPtest]
55 Test the speed of the builtin checksumming functions. If no argument is given,
56 all of them are tested. Or a comma separated list can be passed, in which
57 case the given ones are tested.
59 .BI \-\-cmdhelp \fR=\fPcommand
60 Print help information for \fIcommand\fR. May be `all' for all commands.
62 .BI \-\-enghelp \fR=\fPioengine[,command]
63 List all commands defined by \fIioengine\fR, or print help for \fIcommand\fR defined by \fIioengine\fR.
65 .BI \-\-showcmd \fR=\fPjobfile
66 Convert \fIjobfile\fR to a set of command-line options.
68 .BI \-\-eta \fR=\fPwhen
69 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
70 be one of `always', `never' or `auto'.
72 .BI \-\-eta\-newline \fR=\fPtime
73 Force an ETA newline for every `time` period passed.
75 .BI \-\-status\-interval \fR=\fPtime
76 Report full output status every `time` period passed.
79 Turn on safety read-only checks, preventing any attempted write.
81 .BI \-\-section \fR=\fPsec
82 Only run section \fIsec\fR from job file. This option can be used multiple times to add more sections to run.
84 .BI \-\-alloc\-size \fR=\fPkb
85 Set the internal smalloc pool size to \fIkb\fP kilobytes.
87 .BI \-\-warnings\-fatal
88 All fio parser warnings are fatal, causing fio to exit with an error.
90 .BI \-\-max\-jobs \fR=\fPnr
91 Set the maximum allowed number of jobs (threads/processes) to support.
93 .BI \-\-server \fR=\fPargs
94 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
96 .BI \-\-daemonize \fR=\fPpidfile
97 Background a fio server, writing the pid to the given pid file.
99 .BI \-\-client \fR=\fPhost
100 Instead of running the jobs locally, send and run them on the given host or set of hosts. See client/server section.
102 .BI \-\-idle\-prof \fR=\fPoption
103 Report cpu idleness on a system or percpu basis (\fIoption\fP=system,percpu) or run unit work calibration only (\fIoption\fP=calibrate).
104 .SH "JOB FILE FORMAT"
105 Job files are in `ini' format. They consist of one or more
106 job definitions, which begin with a job name in square brackets and
107 extend to the next job name. The job name can be any ASCII string
108 except `global', which has a special meaning. Following the job name is
109 a sequence of zero or more parameters, one per line, that define the
110 behavior of the job. Any line starting with a `;' or `#' character is
111 considered a comment and ignored.
113 If \fIjobfile\fR is specified as `-', the job file will be read from
116 The global section contains default parameters for jobs specified in the
117 job file. A job is only affected by global sections residing above it,
118 and there may be any number of global sections. Specific job definitions
119 may override any parameter set in global sections.
122 Some parameters may take arguments of a specific type.
123 Anywhere a numeric value is required, an arithmetic expression may be used,
124 provided it is surrounded by parentheses. Supported operators are:
132 .B multiplication (*)
138 .B exponentiation (^)
142 For time values in expressions, units are microseconds by default. This is
143 different than for time values not in expressions (not enclosed in
144 parentheses). The types used are:
147 String: a sequence of alphanumeric characters.
150 SI integer: a whole number, possibly containing a suffix denoting the base unit
151 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
152 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
153 respectively. If prefixed with '0x', the value is assumed to be base 16
154 (hexadecimal). A suffix may include a trailing 'b', for instance 'kb' is
155 identical to 'k'. You can specify a base 10 value by using 'KiB', 'MiB','GiB',
156 etc. This is useful for disk drives where values are often given in base 10
157 values. Specifying '30GiB' will get you 30*1000^3 bytes.
158 When specifying times the default suffix meaning changes, still denoting the
159 base unit of the value, but accepted suffixes are 'D' (days), 'H' (hours), 'M'
160 (minutes), 'S' Seconds, 'ms' (or msec) milli seconds, 'us' (or 'usec') micro
161 seconds. Time values without a unit specify seconds.
162 The suffixes are not case sensitive.
165 Boolean: a true or false value. `0' denotes false, `1' denotes true.
168 Integer range: a range of integers specified in the format
169 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
170 \fIupper\fR may contain a suffix as described above. If an option allows two
171 sets of ranges, they are separated with a `,' or `/' character. For example:
175 List of floating numbers: A list of floating numbers, separated by
180 May be used to override the job name. On the command line, this parameter
181 has the special purpose of signalling the start of a new job.
183 .BI wait_for \fR=\fPstr
184 Specifies the name of the already defined job to wait for. Single waitee name
185 only may be specified. If set, the job won't be started until all workers of
186 the waitee job are done. Wait_for operates on the job name basis, so there are
187 a few limitations. First, the waitee must be defined prior to the waiter job
188 (meaning no forward references). Second, if a job is being referenced as a
189 waitee, it must have a unique name (no duplicate waitees).
191 .BI description \fR=\fPstr
192 Human-readable description of the job. It is printed when the job is run, but
193 otherwise has no special purpose.
195 .BI directory \fR=\fPstr
196 Prefix filenames with this directory. Used to place files in a location other
198 You can specify a number of directories by separating the names with a ':'
199 character. These directories will be assigned equally distributed to job clones
200 creates with \fInumjobs\fR as long as they are using generated filenames.
201 If specific \fIfilename(s)\fR are set fio will use the first listed directory,
202 and thereby matching the \fIfilename\fR semantic which generates a file each
203 clone if not specified, but let all clones use the same if set. See
204 \fIfilename\fR for considerations regarding escaping certain characters on
207 .BI filename \fR=\fPstr
209 normally makes up a file name based on the job name, thread number, and file
210 number. If you want to share files between threads in a job or several jobs,
211 specify a \fIfilename\fR for each of them to override the default.
212 If the I/O engine is file-based, you can specify
213 a number of files by separating the names with a `:' character. `\-' is a
214 reserved name, meaning stdin or stdout, depending on the read/write direction
215 set. On Windows, disk devices are accessed as \\.\PhysicalDrive0 for the first
216 device, \\.\PhysicalDrive1 for the second etc. Note: Windows and FreeBSD
217 prevent write access to areas of the disk containing in-use data
218 (e.g. filesystems). If the wanted filename does need to include a colon, then
219 escape that with a '\\' character. For instance, if the filename is
220 "/dev/dsk/foo@3,0:c", then you would use filename="/dev/dsk/foo@3,0\\:c".
222 .BI filename_format \fR=\fPstr
223 If sharing multiple files between jobs, it is usually necessary to have
224 fio generate the exact names that you want. By default, fio will name a file
225 based on the default file format specification of
226 \fBjobname.jobnumber.filenumber\fP. With this option, that can be
227 customized. Fio will recognize and replace the following keywords in this
233 The name of the worker thread or process.
236 The incremental number of the worker thread or process.
239 The incremental number of the file for that worker thread or process.
242 To have dependent jobs share a set of files, this option can be set to
243 have fio generate filenames that are shared between the two. For instance,
244 if \fBtestfiles.$filenum\fR is specified, file number 4 for any job will
245 be named \fBtestfiles.4\fR. The default of \fB$jobname.$jobnum.$filenum\fR
246 will be used if no other format specifier is given.
250 .BI lockfile \fR=\fPstr
251 Fio defaults to not locking any files before it does IO to them. If a file or
252 file descriptor is shared, fio can serialize IO to that file to make the end
253 result consistent. This is usual for emulating real workloads that share files.
259 No locking. This is the default.
262 Only one thread or process may do IO at a time, excluding all others.
265 Read-write locking on the file. Many readers may access the file at the same
266 time, but writes get exclusive access.
270 .BI opendir \fR=\fPstr
271 Recursively open any files below directory \fIstr\fR.
273 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
274 Type of I/O pattern. Accepted values are:
285 Sequential trim (Linux block devices only).
294 Random trim (Linux block devices only).
297 Mixed sequential reads and writes.
300 Mixed random reads and writes.
303 Trim and write mixed workload. Blocks will be trimmed first, then the same
304 blocks will be written to.
307 For mixed I/O, the default split is 50/50. For certain types of io the result
308 may still be skewed a bit, since the speed may be different. It is possible to
309 specify a number of IO's to do before getting a new offset, this is done by
310 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
311 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
312 value of 8. If the postfix is used with a sequential IO pattern, then the value
313 specified will be added to the generated offset for each IO. For instance,
314 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
315 into sequential IO with holes. See the \fBrw_sequencer\fR option.
318 .BI rw_sequencer \fR=\fPstr
319 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
320 then this option controls how that number modifies the IO offset being
321 generated. Accepted values are:
326 Generate sequential offset
329 Generate the same offset
332 \fBsequential\fR is only useful for random IO, where fio would normally
333 generate a new random offset for every IO. If you append eg 8 to randread, you
334 would get a new random offset for every 8 IO's. The result would be a seek for
335 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
336 that. As sequential IO is already sequential, setting \fBsequential\fR for that
337 would not result in any differences. \fBidentical\fR behaves in a similar
338 fashion, except it sends the same offset 8 number of times before generating a
343 .BI kb_base \fR=\fPint
344 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
345 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
346 reasons. Allowed values are 1024 or 1000, with 1024 being the default.
348 .BI unified_rw_reporting \fR=\fPbool
349 Fio normally reports statistics on a per data direction basis, meaning that
350 read, write, and trim are accounted and reported separately. If this option is
351 set fio sums the results and reports them as "mixed" instead.
353 .BI randrepeat \fR=\fPbool
354 Seed the random number generator used for random I/O patterns in a predictable
355 way so the pattern is repeatable across runs. Default: true.
357 .BI allrandrepeat \fR=\fPbool
358 Seed all random number generators in a predictable way so results are
359 repeatable across runs. Default: false.
361 .BI randseed \fR=\fPint
362 Seed the random number generators based on this seed value, to be able to
363 control what sequence of output is being generated. If not set, the random
364 sequence depends on the \fBrandrepeat\fR setting.
366 .BI fallocate \fR=\fPstr
367 Whether pre-allocation is performed when laying down files. Accepted values
373 Do not pre-allocate space.
376 Pre-allocate via \fBposix_fallocate\fR\|(3).
379 Pre-allocate via \fBfallocate\fR\|(2) with FALLOC_FL_KEEP_SIZE set.
382 Backward-compatible alias for 'none'.
385 Backward-compatible alias for 'posix'.
388 May not be available on all supported platforms. 'keep' is only
389 available on Linux. If using ZFS on Solaris this must be set to 'none'
390 because ZFS doesn't support it. Default: 'posix'.
393 .BI fadvise_hint \fR=\fPbool
394 Use \fBposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
395 are likely to be issued. Default: true.
397 .BI fadvise_stream \fR=\fPint
398 Use \fBposix_fadvise\fR\|(2) to advise the kernel what stream ID the
399 writes issued belong to. Only supported on Linux. Note, this option
400 may change going forward.
403 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
404 been transferred, unless limited by other options (\fBruntime\fR, for instance,
405 or increased/descreased by \fBio_size\fR). Unless \fBnrfiles\fR and
406 \fBfilesize\fR options are given, this amount will be divided between the
407 available files for the job. If not set, fio will use the full size of the
408 given files or devices. If the files do not exist, size must be given. It is
409 also possible to give size as a percentage between 1 and 100. If size=20% is
410 given, fio will use 20% of the full size of the given files or devices.
412 .BI io_size \fR=\fPint "\fR,\fB io_limit \fR=\fPint
413 Normally fio operates within the region set by \fBsize\fR, which means that
414 the \fBsize\fR option sets both the region and size of IO to be performed.
415 Sometimes that is not what you want. With this option, it is possible to
416 define just the amount of IO that fio should do. For instance, if \fBsize\fR
417 is set to 20G and \fBio_limit\fR is set to 5G, fio will perform IO within
418 the first 20G but exit when 5G have been done. The opposite is also
419 possible - if \fBsize\fR is set to 20G, and \fBio_size\fR is set to 40G, then
420 fio will do 40G of IO within the 0..20G region.
422 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
423 Sets size to something really large and waits for ENOSPC (no space left on
424 device) as the terminating condition. Only makes sense with sequential write.
425 For a read workload, the mount point will be filled first then IO started on
426 the result. This option doesn't make sense if operating on a raw device node,
427 since the size of that is already known by the file system. Additionally,
428 writing beyond end-of-device will not return ENOSPC there.
430 .BI filesize \fR=\fPirange
431 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
432 for files at random within the given range, limited to \fBsize\fR in total (if
433 that is given). If \fBfilesize\fR is not specified, each created file is the
436 .BI file_append \fR=\fPbool
437 Perform IO after the end of the file. Normally fio will operate within the
438 size of a file. If this option is set, then fio will append to the file
439 instead. This has identical behavior to setting \fRoffset\fP to the size
440 of a file. This option is ignored on non-regular files.
442 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
443 Block size for I/O units. Default: 4k. Values for reads, writes, and trims
444 can be specified separately in the format \fIread\fR,\fIwrite\fR,\fItrim\fR
445 either of which may be empty to leave that value at its default. If a trailing
446 comma isn't given, the remainder will inherit the last value set.
448 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
449 Specify a range of I/O block sizes. The issued I/O unit will always be a
450 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
451 to both reads and writes if only one range is given, but can be specified
452 separately with a comma separating the values. Example: bsrange=1k-4k,2k-8k.
453 Also (see \fBblocksize\fR).
455 .BI bssplit \fR=\fPstr
456 This option allows even finer grained control of the block sizes issued,
457 not just even splits between them. With this option, you can weight various
458 block sizes for exact control of the issued IO for a job that has mixed
459 block sizes. The format of the option is bssplit=blocksize/percentage,
460 optionally adding as many definitions as needed separated by a colon.
461 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
462 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
463 splits to reads and writes. The format is identical to what the
464 \fBbs\fR option accepts, the read and write parts are separated with a
467 .B blocksize_unaligned\fR,\fP bs_unaligned
468 If set, any size in \fBblocksize_range\fR may be used. This typically won't
469 work with direct I/O, as that normally requires sector alignment.
471 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
472 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
473 the minimum blocksize given. Minimum alignment is typically 512b
474 for using direct IO, though it usually depends on the hardware block size.
475 This option is mutually exclusive with using a random map for files, so it
476 will turn off that option.
478 .BI bs_is_seq_rand \fR=\fPbool
479 If this option is set, fio will use the normal read,write blocksize settings as
480 sequential,random instead. Any random read or write will use the WRITE
481 blocksize settings, and any sequential read or write will use the READ
485 Initialize buffers with all zeros. Default: fill buffers with random data.
488 If this option is given, fio will refill the IO buffers on every submit. The
489 default is to only fill it at init time and reuse that data. Only makes sense
490 if zero_buffers isn't specified, naturally. If data verification is enabled,
491 refill_buffers is also automatically enabled.
493 .BI scramble_buffers \fR=\fPbool
494 If \fBrefill_buffers\fR is too costly and the target is using data
495 deduplication, then setting this option will slightly modify the IO buffer
496 contents to defeat normal de-dupe attempts. This is not enough to defeat
497 more clever block compression attempts, but it will stop naive dedupe
498 of blocks. Default: true.
500 .BI buffer_compress_percentage \fR=\fPint
501 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
502 that compress to the specified level. Fio does this by providing a mix of
503 random data and a fixed pattern. The fixed pattern is either zeroes, or the
504 pattern specified by \fBbuffer_pattern\fR. If the pattern option is used, it
505 might skew the compression ratio slightly. Note that this is per block size
506 unit, for file/disk wide compression level that matches this setting. Note
507 that this is per block size unit, for file/disk wide compression level that
508 matches this setting, you'll also want to set refill_buffers.
510 .BI buffer_compress_chunk \fR=\fPint
511 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
512 big the ranges of random data and zeroed data is. Without this set, fio will
513 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
514 the remaining zeroed. With this set to some chunk size smaller than the block
515 size, fio can alternate random and zeroed data throughout the IO buffer.
517 .BI buffer_pattern \fR=\fPstr
518 If set, fio will fill the IO buffers with this pattern. If not set, the contents
519 of IO buffers is defined by the other options related to buffer contents. The
520 setting can be any pattern of bytes, and can be prefixed with 0x for hex
521 values. It may also be a string, where the string must then be wrapped with
525 \fBbuffer_pattern\fR="abcd"
529 \fBbuffer_pattern\fR=-12
533 \fBbuffer_pattern\fR=0xdeadface
536 Also you can combine everything together in any order:
539 \fBbuffer_pattern\fR=0xdeadface"abcd"-12
543 .BI dedupe_percentage \fR=\fPint
544 If set, fio will generate this percentage of identical buffers when writing.
545 These buffers will be naturally dedupable. The contents of the buffers depend
546 on what other buffer compression settings have been set. It's possible to have
547 the individual buffers either fully compressible, or not at all. This option
548 only controls the distribution of unique buffers.
550 .BI nrfiles \fR=\fPint
551 Number of files to use for this job. Default: 1.
553 .BI openfiles \fR=\fPint
554 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
556 .BI file_service_type \fR=\fPstr
557 Defines how files to service are selected. The following types are defined:
562 Choose a file at random.
565 Round robin over opened files (default).
568 Do each file in the set sequentially.
571 The number of I/Os to issue before switching to a new file can be specified by
572 appending `:\fIint\fR' to the service type.
575 .BI ioengine \fR=\fPstr
576 Defines how the job issues I/O. The following types are defined:
581 Basic \fBread\fR\|(2) or \fBwrite\fR\|(2) I/O. \fBfseek\fR\|(2) is used to
582 position the I/O location.
585 Basic \fBpread\fR\|(2) or \fBpwrite\fR\|(2) I/O.
588 Basic \fBreadv\fR\|(2) or \fBwritev\fR\|(2) I/O. Will emulate queuing by
589 coalescing adjacent IOs into a single submission.
592 Basic \fBpreadv\fR\|(2) or \fBpwritev\fR\|(2) I/O.
595 Linux native asynchronous I/O. This ioengine defines engine specific options.
598 POSIX asynchronous I/O using \fBaio_read\fR\|(3) and \fBaio_write\fR\|(3).
601 Solaris native asynchronous I/O.
604 Windows native asynchronous I/O.
607 File is memory mapped with \fBmmap\fR\|(2) and data copied using
611 \fBsplice\fR\|(2) is used to transfer the data and \fBvmsplice\fR\|(2) to
612 transfer data from user-space to the kernel.
615 Use the syslet system calls to make regular read/write asynchronous.
618 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
619 the target is an sg character device, we use \fBread\fR\|(2) and
620 \fBwrite\fR\|(2) for asynchronous I/O.
623 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
624 itself and for debugging and testing purposes.
627 Transfer over the network. The protocol to be used can be defined with the
628 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
629 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
630 This ioengine defines engine specific options.
633 Like \fBnet\fR, but uses \fBsplice\fR\|(2) and \fBvmsplice\fR\|(2) to map data
634 and send/receive. This ioengine defines engine specific options.
637 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
638 \fBcpucycles\fR parameters.
641 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
642 approach to asynchronous I/O.
644 See <http://www.xmailserver.org/guasi\-lib.html>.
647 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
648 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
651 Loads an external I/O engine object file. Append the engine filename as
655 IO engine that does regular linux native fallocate call to simulate data
656 transfer as fio ioengine
658 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
660 DIR_WRITE does fallocate(,mode = 0)
662 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
665 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
666 request to DDIR_WRITE event
669 IO engine supporting direct access to Ceph Rados Block Devices (RBD) via librbd
670 without the need to use the kernel rbd driver. This ioengine defines engine specific
674 Using Glusterfs libgfapi sync interface to direct access to Glusterfs volumes without
675 having to go through FUSE. This ioengine defines engine specific
679 Using Glusterfs libgfapi async interface to direct access to Glusterfs volumes without
680 having to go through FUSE. This ioengine defines engine specific
684 Read and write through Hadoop (HDFS). The \fBfilename\fR option is used to
685 specify host,port of the hdfs name-node to connect. This engine interprets
686 offsets a little differently. In HDFS, files once created cannot be modified.
687 So random writes are not possible. To imitate this, libhdfs engine expects
688 bunch of small files to be created over HDFS, and engine will randomly pick a
689 file out of those files based on the offset generated by fio backend. (see the
690 example job file to create such files, use rw=write option). Please note, you
691 might want to set necessary environment variables to work with hdfs/libhdfs
695 Read, write and erase an MTD character device (e.g., /dev/mtd0). Discards are
696 treated as erases. Depending on the underlying device type, the I/O may have
697 to go in a certain pattern, e.g., on NAND, writing sequentially to erase blocks
698 and discarding before overwriting. The writetrim mode works well for this
704 .BI iodepth \fR=\fPint
705 Number of I/O units to keep in flight against the file. Note that increasing
706 iodepth beyond 1 will not affect synchronous ioengines (except for small
707 degress when verify_async is in use). Even async engines may impose OS
708 restrictions causing the desired depth not to be achieved. This may happen on
709 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
710 not async on that OS. Keep an eye on the IO depth distribution in the
711 fio output to verify that the achieved depth is as expected. Default: 1.
713 .BI iodepth_batch \fR=\fPint "\fR,\fP iodepth_batch_submit" \fR=\fPint
714 This defines how many pieces of IO to submit at once. It defaults to 1
715 which means that we submit each IO as soon as it is available, but can
716 be raised to submit bigger batches of IO at the time. If it is set to 0
717 the \fBiodepth\fR value will be used.
719 .BI iodepth_batch_complete_min \fR=\fPint "\fR,\fP iodepth_batch_complete" \fR=\fPint
720 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
721 means that we'll ask for a minimum of 1 IO in the retrieval process from the
722 kernel. The IO retrieval will go on until we hit the limit set by
723 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
724 completed events before queuing more IO. This helps reduce IO latency, at the
725 cost of more retrieval system calls.
727 .BI iodepth_batch_complete_max \fR=\fPint
728 This defines maximum pieces of IO to
729 retrieve at once. This variable should be used along with
730 \fBiodepth_batch_complete_min\fR=int variable, specifying the range
731 of min and max amount of IO which should be retrieved. By default
732 it is equal to \fBiodepth_batch_complete_min\fR value.
737 \fBiodepth_batch_complete_min\fR=1
739 \fBiodepth_batch_complete_max\fR=<iodepth>
742 which means that we will retrieve at leat 1 IO and up to the
743 whole submitted queue depth. If none of IO has been completed
748 \fBiodepth_batch_complete_min\fR=0
750 \fBiodepth_batch_complete_max\fR=<iodepth>
753 which means that we can retrieve up to the whole submitted
754 queue depth, but if none of IO has been completed yet, we will
755 NOT wait and immediately exit the system call. In this example
756 we simply do polling.
759 .BI iodepth_low \fR=\fPint
760 Low watermark indicating when to start filling the queue again. Default:
763 .BI io_submit_mode \fR=\fPstr
764 This option controls how fio submits the IO to the IO engine. The default is
765 \fBinline\fR, which means that the fio job threads submit and reap IO directly.
766 If set to \fBoffload\fR, the job threads will offload IO submission to a
767 dedicated pool of IO threads. This requires some coordination and thus has a
768 bit of extra overhead, especially for lower queue depth IO where it can
769 increase latencies. The benefit is that fio can manage submission rates
770 independently of the device completion rates. This avoids skewed latency
771 reporting if IO gets back up on the device side (the coordinated omission
774 .BI direct \fR=\fPbool
775 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
777 .BI atomic \fR=\fPbool
778 If value is true, attempt to use atomic direct IO. Atomic writes are guaranteed
779 to be stable once acknowledged by the operating system. Only Linux supports
782 .BI buffered \fR=\fPbool
783 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
786 .BI offset \fR=\fPint
787 Offset in the file to start I/O. Data before the offset will not be touched.
789 .BI offset_increment \fR=\fPint
790 If this is provided, then the real offset becomes the
791 offset + offset_increment * thread_number, where the thread number is a
792 counter that starts at 0 and is incremented for each sub-job (i.e. when
793 numjobs option is specified). This option is useful if there are several jobs
794 which are intended to operate on a file in parallel disjoint segments, with
795 even spacing between the starting points.
797 .BI number_ios \fR=\fPint
798 Fio will normally perform IOs until it has exhausted the size of the region
799 set by \fBsize\fR, or if it exhaust the allocated time (or hits an error
800 condition). With this setting, the range/size can be set independently of
801 the number of IOs to perform. When fio reaches this number, it will exit
802 normally and report status. Note that this does not extend the amount
803 of IO that will be done, it will only stop fio if this condition is met
804 before other end-of-job criteria.
807 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
808 0, don't sync. Default: 0.
810 .BI fdatasync \fR=\fPint
811 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
812 data parts of the file. Default: 0.
814 .BI write_barrier \fR=\fPint
815 Make every Nth write a barrier write.
817 .BI sync_file_range \fR=\fPstr:int
818 Use \fBsync_file_range\fR\|(2) for every \fRval\fP number of write operations. Fio will
819 track range of writes that have happened since the last \fBsync_file_range\fR\|(2) call.
820 \fRstr\fP can currently be one or more of:
824 SYNC_FILE_RANGE_WAIT_BEFORE
827 SYNC_FILE_RANGE_WRITE
830 SYNC_FILE_RANGE_WRITE
834 So if you do sync_file_range=wait_before,write:8, fio would use
835 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
836 Also see the \fBsync_file_range\fR\|(2) man page. This option is Linux specific.
838 .BI overwrite \fR=\fPbool
839 If writing, setup the file first and do overwrites. Default: false.
841 .BI end_fsync \fR=\fPbool
842 Sync file contents when a write stage has completed. Default: false.
844 .BI fsync_on_close \fR=\fPbool
845 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
846 it will happen on every close, not just at the end of the job. Default: false.
848 .BI rwmixread \fR=\fPint
849 Percentage of a mixed workload that should be reads. Default: 50.
851 .BI rwmixwrite \fR=\fPint
852 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
853 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
854 overrides the first. This may interfere with a given rate setting, if fio is
855 asked to limit reads or writes to a certain rate. If that is the case, then
856 the distribution may be skewed. Default: 50.
858 .BI random_distribution \fR=\fPstr:float
859 By default, fio will use a completely uniform random distribution when asked
860 to perform random IO. Sometimes it is useful to skew the distribution in
861 specific ways, ensuring that some parts of the data is more hot than others.
862 Fio includes the following distribution models:
866 Uniform random distribution
876 When using a zipf or pareto distribution, an input value is also needed to
877 define the access pattern. For zipf, this is the zipf theta. For pareto,
878 it's the pareto power. Fio includes a test program, genzipf, that can be
879 used visualize what the given input values will yield in terms of hit rates.
880 If you wanted to use zipf with a theta of 1.2, you would use
881 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
882 fio will disable use of the random map.
884 .BI percentage_random \fR=\fPint
885 For a random workload, set how big a percentage should be random. This defaults
886 to 100%, in which case the workload is fully random. It can be set from
887 anywhere from 0 to 100. Setting it to 0 would make the workload fully
888 sequential. It is possible to set different values for reads, writes, and
889 trim. To do so, simply use a comma separated list. See \fBblocksize\fR.
892 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
893 this parameter is given, a new offset will be chosen without looking at past
894 I/O history. This parameter is mutually exclusive with \fBverify\fR.
896 .BI softrandommap \fR=\fPbool
897 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
898 fails to allocate the map, if this option is set it will continue without a
899 random block map. As coverage will not be as complete as with random maps, this
900 option is disabled by default.
902 .BI random_generator \fR=\fPstr
903 Fio supports the following engines for generating IO offsets for random IO:
907 Strong 2^88 cycle random number generator
910 Linear feedback shift register generator
913 Strong 64-bit 2^258 cycle random number generator
917 Tausworthe is a strong random number generator, but it requires tracking on the
918 side if we want to ensure that blocks are only read or written once. LFSR
919 guarantees that we never generate the same offset twice, and it's also less
920 computationally expensive. It's not a true random generator, however, though
921 for IO purposes it's typically good enough. LFSR only works with single block
922 sizes, not with workloads that use multiple block sizes. If used with such a
923 workload, fio may read or write some blocks multiple times.
926 Run job with given nice value. See \fBnice\fR\|(2).
929 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
932 .BI prioclass \fR=\fPint
933 Set I/O priority class. See \fBionice\fR\|(1).
935 .BI thinktime \fR=\fPint
936 Stall job for given number of microseconds between issuing I/Os.
938 .BI thinktime_spin \fR=\fPint
939 Pretend to spend CPU time for given number of microseconds, sleeping the rest
940 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
942 .BI thinktime_blocks \fR=\fPint
943 Only valid if thinktime is set - control how many blocks to issue, before
944 waiting \fBthinktime\fR microseconds. If not set, defaults to 1 which will
945 make fio wait \fBthinktime\fR microseconds after every block. This
946 effectively makes any queue depth setting redundant, since no more than 1 IO
947 will be queued before we have to complete it and do our thinktime. In other
948 words, this setting effectively caps the queue depth if the latter is larger.
952 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
953 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
954 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
955 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
956 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
957 limit writes (to 500KB/sec), the latter will only limit reads.
959 .BI rate_min \fR=\fPint
960 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
961 Failing to meet this requirement will cause the job to exit. The same format
962 as \fBrate\fR is used for read vs write separation.
964 .BI rate_iops \fR=\fPint
965 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
966 specified independently of bandwidth. The same format as \fBrate\fR is used for
967 read vs write separation. If \fBblocksize\fR is a range, the smallest block
968 size is used as the metric.
970 .BI rate_iops_min \fR=\fPint
971 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
972 is used for read vs write separation.
974 .BI rate_process \fR=\fPstr
975 This option controls how fio manages rated IO submissions. The default is
976 \fBlinear\fR, which submits IO in a linear fashion with fixed delays between
977 IOs that gets adjusted based on IO completion rates. If this is set to
978 \fBpoisson\fR, fio will submit IO based on a more real world random request
979 flow, known as the Poisson process
980 (https://en.wikipedia.org/wiki/Poisson_process). The lambda will be
981 10^6 / IOPS for the given workload.
983 .BI rate_cycle \fR=\fPint
984 Average bandwidth for \fBrate\fR and \fBrate_min\fR over this number of
985 milliseconds. Default: 1000ms.
987 .BI latency_target \fR=\fPint
988 If set, fio will attempt to find the max performance point that the given
989 workload will run at while maintaining a latency below this target. The
990 values is given in microseconds. See \fBlatency_window\fR and
991 \fBlatency_percentile\fR.
993 .BI latency_window \fR=\fPint
994 Used with \fBlatency_target\fR to specify the sample window that the job
995 is run at varying queue depths to test the performance. The value is given
998 .BI latency_percentile \fR=\fPfloat
999 The percentage of IOs that must fall within the criteria specified by
1000 \fBlatency_target\fR and \fBlatency_window\fR. If not set, this defaults
1001 to 100.0, meaning that all IOs must be equal or below to the value set
1002 by \fBlatency_target\fR.
1004 .BI max_latency \fR=\fPint
1005 If set, fio will exit the job if it exceeds this maximum latency. It will exit
1006 with an ETIME error.
1008 .BI cpumask \fR=\fPint
1009 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
1010 may run on. See \fBsched_setaffinity\fR\|(2).
1012 .BI cpus_allowed \fR=\fPstr
1013 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
1015 .BI cpus_allowed_policy \fR=\fPstr
1016 Set the policy of how fio distributes the CPUs specified by \fBcpus_allowed\fR
1017 or \fBcpumask\fR. Two policies are supported:
1022 All jobs will share the CPU set specified.
1025 Each job will get a unique CPU from the CPU set.
1028 \fBshared\fR is the default behaviour, if the option isn't specified. If
1029 \fBsplit\fR is specified, then fio will assign one cpu per job. If not enough
1030 CPUs are given for the jobs listed, then fio will roundrobin the CPUs in
1035 .BI numa_cpu_nodes \fR=\fPstr
1036 Set this job running on specified NUMA nodes' CPUs. The arguments allow
1037 comma delimited list of cpu numbers, A-B ranges, or 'all'.
1039 .BI numa_mem_policy \fR=\fPstr
1040 Set this job's memory policy and corresponding NUMA nodes. Format of
1044 .B <mode>[:<nodelist>]
1047 is one of the following memory policy:
1049 .B default, prefer, bind, interleave, local
1052 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
1053 needed to be specified. For \fBprefer\fR, only one node is
1054 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
1055 comma delimited list of numbers, A-B ranges, or 'all'.
1057 .BI startdelay \fR=\fPirange
1058 Delay start of job for the specified number of seconds. Supports all time
1059 suffixes to allow specification of hours, minutes, seconds and
1060 milliseconds - seconds are the default if a unit is omitted.
1061 Can be given as a range which causes each thread to choose randomly out of the
1064 .BI runtime \fR=\fPint
1065 Terminate processing after the specified number of seconds.
1068 If given, run for the specified \fBruntime\fR duration even if the files are
1069 completely read or written. The same workload will be repeated as many times
1070 as \fBruntime\fR allows.
1072 .BI ramp_time \fR=\fPint
1073 If set, fio will run the specified workload for this amount of time before
1074 logging any performance numbers. Useful for letting performance settle before
1075 logging results, thus minimizing the runtime required for stable results. Note
1076 that the \fBramp_time\fR is considered lead in time for a job, thus it will
1077 increase the total runtime if a special timeout or runtime is specified.
1079 .BI invalidate \fR=\fPbool
1080 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
1082 .BI sync \fR=\fPbool
1083 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1084 this means using O_SYNC. Default: false.
1086 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
1087 Allocation method for I/O unit buffer. Allowed values are:
1092 Allocate memory with \fBmalloc\fR\|(3).
1095 Use shared memory buffers allocated through \fBshmget\fR\|(2).
1098 Same as \fBshm\fR, but use huge pages as backing.
1101 Use \fBmmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
1102 is given after the option in the format `:\fIfile\fR'.
1105 Same as \fBmmap\fR, but use huge files as backing.
1108 Same as \fBmmap\fR, but use a MMAP_SHARED mapping.
1111 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
1112 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
1113 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
1114 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
1115 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
1116 and the documentation for that. Normally you just need to echo an appropriate
1117 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
1121 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
1122 This indicates the memory alignment of the IO memory buffers. Note that the
1123 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
1124 the alignment of the following buffers are given by the \fBbs\fR used. In
1125 other words, if using a \fBbs\fR that is a multiple of the page sized in the
1126 system, all buffers will be aligned to this value. If using a \fBbs\fR that
1127 is not page aligned, the alignment of subsequent IO memory buffers is the
1128 sum of the \fBiomem_align\fR and \fBbs\fR used.
1130 .BI hugepage\-size \fR=\fPint
1131 Defines the size of a huge page. Must be at least equal to the system setting.
1132 Should be a multiple of 1MB. Default: 4MB.
1135 Terminate all jobs when one finishes. Default: wait for each job to finish.
1137 .B exitall_on_error \fR=\fPbool
1138 Terminate all jobs if one job finishes in error. Default: wait for each job
1141 .BI bwavgtime \fR=\fPint
1142 Average bandwidth calculations over the given time in milliseconds. Default:
1145 .BI iopsavgtime \fR=\fPint
1146 Average IOPS calculations over the given time in milliseconds. Default:
1149 .BI create_serialize \fR=\fPbool
1150 If true, serialize file creation for the jobs. Default: true.
1152 .BI create_fsync \fR=\fPbool
1153 \fBfsync\fR\|(2) data file after creation. Default: true.
1155 .BI create_on_open \fR=\fPbool
1156 If true, the files are not created until they are opened for IO by the job.
1158 .BI create_only \fR=\fPbool
1159 If true, fio will only run the setup phase of the job. If files need to be
1160 laid out or updated on disk, only that will be done. The actual job contents
1163 .BI allow_file_create \fR=\fPbool
1164 If true, fio is permitted to create files as part of its workload. This is
1165 the default behavior. If this option is false, then fio will error out if the
1166 files it needs to use don't already exist. Default: true.
1168 .BI allow_mounted_write \fR=\fPbool
1169 If this isn't set, fio will abort jobs that are destructive (eg that write)
1170 to what appears to be a mounted device or partition. This should help catch
1171 creating inadvertently destructive tests, not realizing that the test will
1172 destroy data on the mounted file system. Default: false.
1174 .BI pre_read \fR=\fPbool
1175 If this is given, files will be pre-read into memory before starting the given
1176 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
1177 pointless to pre-read and then drop the cache. This will only work for IO
1178 engines that are seekable, since they allow you to read the same data
1179 multiple times. Thus it will not work on eg network or splice IO.
1181 .BI unlink \fR=\fPbool
1182 Unlink job files when done. Default: false.
1184 .BI loops \fR=\fPint
1185 Specifies the number of iterations (runs of the same workload) of this job.
1188 .BI verify_only \fR=\fPbool
1189 Do not perform the specified workload, only verify data still matches previous
1190 invocation of this workload. This option allows one to check data multiple
1191 times at a later date without overwriting it. This option makes sense only for
1192 workloads that write data, and does not support workloads with the
1193 \fBtime_based\fR option set.
1195 .BI do_verify \fR=\fPbool
1196 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
1199 .BI verify \fR=\fPstr
1200 Method of verifying file contents after each iteration of the job. Each
1201 verification method also implies verification of special header, which is
1202 written to the beginning of each block. This header also includes meta
1203 information, like offset of the block, block number, timestamp when block
1204 was written, etc. \fBverify\fR=str can be combined with \fBverify_pattern\fR=str
1205 option. The allowed values are:
1209 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1 xxhash
1210 Store appropriate checksum in the header of each block. crc32c-intel is
1211 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
1212 not supported by the system.
1215 This option is deprecated, since now meta information is included in generic
1216 verification header and meta verification happens by default. For detailed
1217 information see the description of the \fBverify\fR=str setting. This option
1218 is kept because of compatibility's sake with old configurations. Do not use it.
1221 Verify a strict pattern. Normally fio includes a header with some basic
1222 information and checksumming, but if this option is set, only the
1223 specific pattern set with \fBverify_pattern\fR is verified.
1226 Pretend to verify. Used for testing internals.
1229 This option can be used for repeated burn-in tests of a system to make sure
1230 that the written data is also correctly read back. If the data direction given
1231 is a read or random read, fio will assume that it should verify a previously
1232 written file. If the data direction includes any form of write, the verify will
1233 be of the newly written data.
1236 .BI verifysort \fR=\fPbool
1237 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
1238 read them back in a sorted manner. Default: true.
1240 .BI verifysort_nr \fR=\fPint
1241 Pre-load and sort verify blocks for a read workload.
1243 .BI verify_offset \fR=\fPint
1244 Swap the verification header with data somewhere else in the block before
1245 writing. It is swapped back before verifying.
1247 .BI verify_interval \fR=\fPint
1248 Write the verification header for this number of bytes, which should divide
1249 \fBblocksize\fR. Default: \fBblocksize\fR.
1251 .BI verify_pattern \fR=\fPstr
1252 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
1253 with totally random bytes, but sometimes it's interesting to fill with a known
1254 pattern for io verification purposes. Depending on the width of the pattern,
1255 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
1256 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
1257 has to be a hex number that starts with either "0x" or "0X". Use with
1258 \fBverify\fP=str. Also, verify_pattern supports %o format, which means that for
1259 each block offset will be written and then verifyied back, e.g.:
1262 \fBverify_pattern\fR=%o
1264 Or use combination of everything:
1267 \fBverify_pattern\fR=0xff%o"abcd"-21
1271 .BI verify_fatal \fR=\fPbool
1272 If true, exit the job on the first observed verification failure. Default:
1275 .BI verify_dump \fR=\fPbool
1276 If set, dump the contents of both the original data block and the data block we
1277 read off disk to files. This allows later analysis to inspect just what kind of
1278 data corruption occurred. Off by default.
1280 .BI verify_async \fR=\fPint
1281 Fio will normally verify IO inline from the submitting thread. This option
1282 takes an integer describing how many async offload threads to create for IO
1283 verification instead, causing fio to offload the duty of verifying IO contents
1284 to one or more separate threads. If using this offload option, even sync IO
1285 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
1286 allows them to have IO in flight while verifies are running.
1288 .BI verify_async_cpus \fR=\fPstr
1289 Tell fio to set the given CPU affinity on the async IO verification threads.
1290 See \fBcpus_allowed\fP for the format used.
1292 .BI verify_backlog \fR=\fPint
1293 Fio will normally verify the written contents of a job that utilizes verify
1294 once that job has completed. In other words, everything is written then
1295 everything is read back and verified. You may want to verify continually
1296 instead for a variety of reasons. Fio stores the meta data associated with an
1297 IO block in memory, so for large verify workloads, quite a bit of memory would
1298 be used up holding this meta data. If this option is enabled, fio will write
1299 only N blocks before verifying these blocks.
1301 .BI verify_backlog_batch \fR=\fPint
1302 Control how many blocks fio will verify if verify_backlog is set. If not set,
1303 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
1304 read back and verified). If \fBverify_backlog_batch\fR is less than
1305 \fBverify_backlog\fR then not all blocks will be verified, if
1306 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
1307 will be verified more than once.
1309 .BI trim_percentage \fR=\fPint
1310 Number of verify blocks to discard/trim.
1312 .BI trim_verify_zero \fR=\fPbool
1313 Verify that trim/discarded blocks are returned as zeroes.
1315 .BI trim_backlog \fR=\fPint
1316 Trim after this number of blocks are written.
1318 .BI trim_backlog_batch \fR=\fPint
1319 Trim this number of IO blocks.
1321 .BI experimental_verify \fR=\fPbool
1322 Enable experimental verification.
1324 .BI verify_state_save \fR=\fPbool
1325 When a job exits during the write phase of a verify workload, save its
1326 current state. This allows fio to replay up until that point, if the
1327 verify state is loaded for the verify read phase.
1329 .BI verify_state_load \fR=\fPbool
1330 If a verify termination trigger was used, fio stores the current write
1331 state of each thread. This can be used at verification time so that fio
1332 knows how far it should verify. Without this information, fio will run
1333 a full verification pass, according to the settings in the job file used.
1335 .B stonewall "\fR,\fP wait_for_previous"
1336 Wait for preceding jobs in the job file to exit before starting this one.
1337 \fBstonewall\fR implies \fBnew_group\fR.
1340 Start a new reporting group. If not given, all jobs in a file will be part
1341 of the same reporting group, unless separated by a stonewall.
1343 .BI numjobs \fR=\fPint
1344 Number of clones (processes/threads performing the same workload) of this job.
1348 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
1352 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
1353 with \fBfork\fR\|(2).
1355 .BI zonesize \fR=\fPint
1356 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
1358 .BI zonerange \fR=\fPint
1359 Give size of an IO zone. See \fBzoneskip\fR.
1361 .BI zoneskip \fR=\fPint
1362 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
1365 .BI write_iolog \fR=\fPstr
1366 Write the issued I/O patterns to the specified file. Specify a separate file
1367 for each job, otherwise the iologs will be interspersed and the file may be
1370 .BI read_iolog \fR=\fPstr
1371 Replay the I/O patterns contained in the specified file generated by
1372 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
1374 .BI replay_no_stall \fR=\fPint
1375 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1376 attempts to respect timing information between I/Os. Enabling
1377 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
1378 still respecting ordering.
1380 .BI replay_redirect \fR=\fPstr
1381 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1382 is to replay the IOPS onto the major/minor device that each IOP was recorded
1383 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
1384 single specified device regardless of the device it was recorded from.
1386 .BI replay_align \fR=\fPint
1387 Force alignment of IO offsets and lengths in a trace to this power of 2 value.
1389 .BI replay_scale \fR=\fPint
1390 Scale sector offsets down by this factor when replaying traces.
1392 .BI per_job_logs \fR=\fPbool
1393 If set, this generates bw/clat/iops log with per file private filenames. If
1394 not set, jobs with identical names will share the log filename. Default: true.
1396 .BI write_bw_log \fR=\fPstr
1397 If given, write a bandwidth log of the jobs in this job file. Can be used to
1398 store data of the bandwidth of the jobs in their lifetime. The included
1399 fio_generate_plots script uses gnuplot to turn these text files into nice
1400 graphs. See \fBwrite_lat_log\fR for behaviour of given filename. For this
1401 option, the postfix is _bw.x.log, where x is the index of the job (1..N,
1402 where N is the number of jobs). If \fBper_job_logs\fR is false, then the
1403 filename will not include the job index.
1405 .BI write_lat_log \fR=\fPstr
1406 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
1407 filename is given with this option, the default filename of
1408 "jobname_type.x.log" is used, where x is the index of the job (1..N, where
1409 N is the number of jobs). Even if the filename is given, fio will still
1410 append the type of log. If \fBper_job_logs\fR is false, then the filename will
1411 not include the job index.
1413 .BI write_iops_log \fR=\fPstr
1414 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
1415 option, the default filename of "jobname_type.x.log" is used, where x is the
1416 index of the job (1..N, where N is the number of jobs). Even if the filename
1417 is given, fio will still append the type of log. If \fBper_job_logs\fR is false,
1418 then the filename will not include the job index.
1420 .BI log_avg_msec \fR=\fPint
1421 By default, fio will log an entry in the iops, latency, or bw log for every
1422 IO that completes. When writing to the disk log, that can quickly grow to a
1423 very large size. Setting this option makes fio average the each log entry
1424 over the specified period of time, reducing the resolution of the log.
1427 .BI log_offset \fR=\fPbool
1428 If this is set, the iolog options will include the byte offset for the IO
1429 entry as well as the other data values.
1431 .BI log_compression \fR=\fPint
1432 If this is set, fio will compress the IO logs as it goes, to keep the memory
1433 footprint lower. When a log reaches the specified size, that chunk is removed
1434 and compressed in the background. Given that IO logs are fairly highly
1435 compressible, this yields a nice memory savings for longer runs. The downside
1436 is that the compression will consume some background CPU cycles, so it may
1437 impact the run. This, however, is also true if the logging ends up consuming
1438 most of the system memory. So pick your poison. The IO logs are saved
1439 normally at the end of a run, by decompressing the chunks and storing them
1440 in the specified log file. This feature depends on the availability of zlib.
1442 .BI log_compression_cpus \fR=\fPstr
1443 Define the set of CPUs that are allowed to handle online log compression
1444 for the IO jobs. This can provide better isolation between performance
1445 sensitive jobs, and background compression work.
1447 .BI log_store_compressed \fR=\fPbool
1448 If set, fio will store the log files in a compressed format. They can be
1449 decompressed with fio, using the \fB\-\-inflate-log\fR command line parameter.
1450 The files will be stored with a \fB\.fz\fR suffix.
1452 .BI block_error_percentiles \fR=\fPbool
1453 If set, record errors in trim block-sized units from writes and trims and output
1454 a histogram of how many trims it took to get to errors, and what kind of error
1457 .BI disable_lat \fR=\fPbool
1458 Disable measurements of total latency numbers. Useful only for cutting
1459 back the number of calls to \fBgettimeofday\fR\|(2), as that does impact performance at
1460 really high IOPS rates. Note that to really get rid of a large amount of these
1461 calls, this option must be used with disable_slat and disable_bw as well.
1463 .BI disable_clat \fR=\fPbool
1464 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1466 .BI disable_slat \fR=\fPbool
1467 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1469 .BI disable_bw_measurement \fR=\fPbool
1470 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1472 .BI lockmem \fR=\fPint
1473 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1474 simulate a smaller amount of memory. The amount specified is per worker.
1476 .BI exec_prerun \fR=\fPstr
1477 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1479 Output is redirected in a file called \fBjobname.prerun.txt\fR
1482 .BI exec_postrun \fR=\fPstr
1483 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1485 Output is redirected in a file called \fBjobname.postrun.txt\fR
1488 .BI ioscheduler \fR=\fPstr
1489 Attempt to switch the device hosting the file to the specified I/O scheduler.
1491 .BI disk_util \fR=\fPbool
1492 Generate disk utilization statistics if the platform supports it. Default: true.
1494 .BI clocksource \fR=\fPstr
1495 Use the given clocksource as the base of timing. The supported options are:
1499 \fBgettimeofday\fR\|(2)
1502 \fBclock_gettime\fR\|(2)
1505 Internal CPU clock source
1509 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1510 (and fio is heavy on time calls). Fio will automatically use this clocksource
1511 if it's supported and considered reliable on the system it is running on,
1512 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1513 means supporting TSC Invariant.
1515 .BI gtod_reduce \fR=\fPbool
1516 Enable all of the \fBgettimeofday\fR\|(2) reducing options (disable_clat, disable_slat,
1517 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1518 \fBgettimeofday\fR\|(2) call count. With this option enabled, we only do about 0.4% of
1519 the gtod() calls we would have done if all time keeping was enabled.
1521 .BI gtod_cpu \fR=\fPint
1522 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1523 the current time. Fio (and databases, for instance) are very intensive on
1524 \fBgettimeofday\fR\|(2) calls. With this option, you can set one CPU aside for doing
1525 nothing but logging current time to a shared memory location. Then the other
1526 threads/processes that run IO workloads need only copy that segment, instead of
1527 entering the kernel with a \fBgettimeofday\fR\|(2) call. The CPU set aside for doing
1528 these time calls will be excluded from other uses. Fio will manually clear it
1529 from the CPU mask of other jobs.
1531 .BI ignore_error \fR=\fPstr
1532 Sometimes you want to ignore some errors during test in that case you can specify
1533 error list for each error type.
1535 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1537 errors for given error type is separated with ':'.
1538 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1540 Example: ignore_error=EAGAIN,ENOSPC:122 .
1542 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1544 .BI error_dump \fR=\fPbool
1545 If set dump every error even if it is non fatal, true by default. If disabled
1546 only fatal error will be dumped
1548 .BI profile \fR=\fPstr
1549 Select a specific builtin performance test.
1551 .BI cgroup \fR=\fPstr
1552 Add job to this control group. If it doesn't exist, it will be created.
1553 The system must have a mounted cgroup blkio mount point for this to work. If
1554 your system doesn't have it mounted, you can do so with:
1556 # mount \-t cgroup \-o blkio none /cgroup
1558 .BI cgroup_weight \fR=\fPint
1559 Set the weight of the cgroup to this value. See the documentation that comes
1560 with the kernel, allowed values are in the range of 100..1000.
1562 .BI cgroup_nodelete \fR=\fPbool
1563 Normally fio will delete the cgroups it has created after the job completion.
1564 To override this behavior and to leave cgroups around after the job completion,
1565 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1566 cgroup files after job completion. Default: false
1569 Instead of running as the invoking user, set the user ID to this value before
1570 the thread/process does any work.
1573 Set group ID, see \fBuid\fR.
1575 .BI unit_base \fR=\fPint
1576 Base unit for reporting. Allowed values are:
1580 Use auto-detection (default).
1590 .BI flow_id \fR=\fPint
1591 The ID of the flow. If not specified, it defaults to being a global flow. See
1595 Weight in token-based flow control. If this value is used, then there is a
1596 \fBflow counter\fR which is used to regulate the proportion of activity between
1597 two or more jobs. fio attempts to keep this flow counter near zero. The
1598 \fBflow\fR parameter stands for how much should be added or subtracted to the
1599 flow counter on each iteration of the main I/O loop. That is, if one job has
1600 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1601 1:8 ratio in how much one runs vs the other.
1603 .BI flow_watermark \fR=\fPint
1604 The maximum value that the absolute value of the flow counter is allowed to
1605 reach before the job must wait for a lower value of the counter.
1607 .BI flow_sleep \fR=\fPint
1608 The period of time, in microseconds, to wait after the flow watermark has been
1609 exceeded before retrying operations
1611 .BI clat_percentiles \fR=\fPbool
1612 Enable the reporting of percentiles of completion latencies.
1614 .BI percentile_list \fR=\fPfloat_list
1615 Overwrite the default list of percentiles for completion latencies and the
1616 block error histogram. Each number is a floating number in the range (0,100],
1617 and the maximum length of the list is 20. Use ':' to separate the
1618 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1619 report the values of completion latency below which 99.5% and 99.9% of
1620 the observed latencies fell, respectively.
1621 .SS "Ioengine Parameters List"
1622 Some parameters are only valid when a specific ioengine is in use. These are
1623 used identically to normal parameters, with the caveat that when used on the
1624 command line, they must come after the ioengine.
1626 .BI (cpu)cpuload \fR=\fPint
1627 Attempt to use the specified percentage of CPU cycles.
1629 .BI (cpu)cpuchunks \fR=\fPint
1630 Split the load into cycles of the given time. In microseconds.
1632 .BI (cpu)exit_on_io_done \fR=\fPbool
1633 Detect when IO threads are done, then exit.
1635 .BI (libaio)userspace_reap
1636 Normally, with the libaio engine in use, fio will use
1637 the io_getevents system call to reap newly returned events.
1638 With this flag turned on, the AIO ring will be read directly
1639 from user-space to reap events. The reaping mode is only
1640 enabled when polling for a minimum of 0 events (eg when
1641 iodepth_batch_complete=0).
1643 .BI (net,netsplice)hostname \fR=\fPstr
1644 The host name or IP address to use for TCP or UDP based IO.
1645 If the job is a TCP listener or UDP reader, the hostname is not
1646 used and must be omitted unless it is a valid UDP multicast address.
1648 .BI (net,netsplice)port \fR=\fPint
1649 The TCP or UDP port to bind to or connect to. If this is used with
1650 \fBnumjobs\fR to spawn multiple instances of the same job type, then
1651 this will be the starting port number since fio will use a range of ports.
1653 .BI (net,netsplice)interface \fR=\fPstr
1654 The IP address of the network interface used to send or receive UDP multicast
1657 .BI (net,netsplice)ttl \fR=\fPint
1658 Time-to-live value for outgoing UDP multicast packets. Default: 1
1660 .BI (net,netsplice)nodelay \fR=\fPbool
1661 Set TCP_NODELAY on TCP connections.
1663 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1664 The network protocol to use. Accepted values are:
1669 Transmission control protocol
1672 Transmission control protocol V6
1675 User datagram protocol
1678 User datagram protocol V6
1684 When the protocol is TCP or UDP, the port must also be given,
1685 as well as the hostname if the job is a TCP listener or UDP
1686 reader. For unix sockets, the normal filename option should be
1687 used and the port is invalid.
1690 .BI (net,netsplice)listen
1691 For TCP network connections, tell fio to listen for incoming
1692 connections rather than initiating an outgoing connection. The
1693 hostname must be omitted if this option is used.
1695 .BI (net, pingpong) \fR=\fPbool
1696 Normally a network writer will just continue writing data, and a network reader
1697 will just consume packets. If pingpong=1 is set, a writer will send its normal
1698 payload to the reader, then wait for the reader to send the same payload back.
1699 This allows fio to measure network latencies. The submission and completion
1700 latencies then measure local time spent sending or receiving, and the
1701 completion latency measures how long it took for the other end to receive and
1702 send back. For UDP multicast traffic pingpong=1 should only be set for a single
1703 reader when multiple readers are listening to the same address.
1705 .BI (net, window_size) \fR=\fPint
1706 Set the desired socket buffer size for the connection.
1708 .BI (net, mss) \fR=\fPint
1709 Set the TCP maximum segment size (TCP_MAXSEG).
1711 .BI (e4defrag,donorname) \fR=\fPstr
1712 File will be used as a block donor (swap extents between files)
1714 .BI (e4defrag,inplace) \fR=\fPint
1715 Configure donor file block allocation strategy
1718 Preallocate donor's file on init
1721 allocate space immediately inside defragment event, and free right after event
1724 .BI (rbd)rbdname \fR=\fPstr
1725 Specifies the name of the RBD.
1727 .BI (rbd)pool \fR=\fPstr
1728 Specifies the name of the Ceph pool containing the RBD.
1730 .BI (rbd)clientname \fR=\fPstr
1731 Specifies the username (without the 'client.' prefix) used to access the Ceph cluster.
1733 .BI (mtd)skipbad \fR=\fPbool
1734 Skip operations against known bad blocks.
1736 While running, \fBfio\fR will display the status of the created jobs. For
1740 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1743 The characters in the first set of brackets denote the current status of each
1744 threads. The possible values are:
1750 Setup but not started.
1756 Initialized, waiting.
1759 Running, doing sequential reads.
1762 Running, doing random reads.
1765 Running, doing sequential writes.
1768 Running, doing random writes.
1771 Running, doing mixed sequential reads/writes.
1774 Running, doing mixed random reads/writes.
1777 Running, currently waiting for \fBfsync\fR\|(2).
1780 Running, verifying written data.
1783 Exited, not reaped by main thread.
1786 Exited, thread reaped.
1790 The second set of brackets shows the estimated completion percentage of
1791 the current group. The third set shows the read and write I/O rate,
1792 respectively. Finally, the estimated run time of the job is displayed.
1794 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1795 for each thread, each group of threads, and each disk, in that order.
1797 Per-thread statistics first show the threads client number, group-id, and
1798 error code. The remaining figures are as follows:
1802 Number of megabytes of I/O performed.
1805 Average data rate (bandwidth).
1811 Submission latency minimum, maximum, average and standard deviation. This is
1812 the time it took to submit the I/O.
1815 Completion latency minimum, maximum, average and standard deviation. This
1816 is the time between submission and completion.
1819 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1820 and standard deviation.
1823 CPU usage statistics. Includes user and system time, number of context switches
1824 this thread went through and number of major and minor page faults.
1827 Distribution of I/O depths. Each depth includes everything less than (or equal)
1828 to it, but greater than the previous depth.
1831 Number of read/write requests issued, and number of short read/write requests.
1834 Distribution of I/O completion latencies. The numbers follow the same pattern
1838 The group statistics show:
1843 Number of megabytes I/O performed.
1846 Aggregate bandwidth of threads in the group.
1849 Minimum average bandwidth a thread saw.
1852 Maximum average bandwidth a thread saw.
1855 Shortest runtime of threads in the group.
1858 Longest runtime of threads in the group.
1862 Finally, disk statistics are printed with reads first:
1867 Number of I/Os performed by all groups.
1870 Number of merges in the I/O scheduler.
1873 Number of ticks we kept the disk busy.
1876 Total time spent in the disk queue.
1883 It is also possible to get fio to dump the current output while it is
1884 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1887 If the \fB\-\-minimal\fR / \fB\-\-append-terse\fR options are given, the
1888 results will be printed/appended in a semicolon-delimited format suitable for
1890 A job description (if provided) follows on a new line. Note that the first
1891 number in the line is the version number. If the output has to be changed
1892 for some reason, this number will be incremented by 1 to signify that
1893 change. The fields are:
1896 .B terse version, fio version, jobname, groupid, error
1900 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1904 .B min, max, mean, standard deviation
1908 .B min, max, mean, standard deviation
1910 Completion latency percentiles (20 fields):
1912 .B Xth percentile=usec
1916 .B min, max, mean, standard deviation
1920 .B min, max, aggregate percentage of total, mean, standard deviation
1926 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1930 .B min, max, mean, standard deviation
1934 .B min, max, mean, standard deviation
1936 Completion latency percentiles (20 fields):
1938 .B Xth percentile=usec
1942 .B min, max, mean, standard deviation
1946 .B min, max, aggregate percentage of total, mean, standard deviation
1952 .B user, system, context switches, major page faults, minor page faults
1955 IO depth distribution:
1957 .B <=1, 2, 4, 8, 16, 32, >=64
1960 IO latency distribution:
1964 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1968 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1972 Disk utilization (1 for each disk used):
1974 .B name, read ios, write ios, read merges, write merges, read ticks, write ticks, read in-queue time, write in-queue time, disk utilization percentage
1977 Error Info (dependent on continue_on_error, default off):
1979 .B total # errors, first error code
1982 .B text description (if provided in config - appears on newline)
1985 Normally you would run fio as a stand-alone application on the machine
1986 where the IO workload should be generated. However, it is also possible to
1987 run the frontend and backend of fio separately. This makes it possible to
1988 have a fio server running on the machine(s) where the IO workload should
1989 be running, while controlling it from another machine.
1991 To start the server, you would do:
1993 \fBfio \-\-server=args\fR
1995 on that machine, where args defines what fio listens to. The arguments
1996 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1997 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1998 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1999 listen to (only valid for TCP/IP, not a local socket). Some examples:
2003 Start a fio server, listening on all interfaces on the default port (8765).
2005 2) fio \-\-server=ip:hostname,4444
2007 Start a fio server, listening on IP belonging to hostname and on port 4444.
2009 3) fio \-\-server=ip6:::1,4444
2011 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
2013 4) fio \-\-server=,4444
2015 Start a fio server, listening on all interfaces on port 4444.
2017 5) fio \-\-server=1.2.3.4
2019 Start a fio server, listening on IP 1.2.3.4 on the default port.
2021 6) fio \-\-server=sock:/tmp/fio.sock
2023 Start a fio server, listening on the local socket /tmp/fio.sock.
2025 When a server is running, you can connect to it from a client. The client
2028 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
2030 where \-\-local-args are arguments that are local to the client where it is
2031 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
2032 are sent to the server. The 'server' string follows the same format as it
2033 does on the server side, to allow IP/hostname/socket and port strings.
2034 You can connect to multiple clients as well, to do that you could run:
2036 fio \-\-client=server2 \-\-client=server2 <job file(s)>
2038 If the job file is located on the fio server, then you can tell the server
2039 to load a local file as well. This is done by using \-\-remote-config:
2041 fio \-\-client=server \-\-remote-config /path/to/file.fio
2043 Then fio will open this local (to the server) job file instead
2044 of being passed one from the client.
2046 If you have many servers (example: 100 VMs/containers), you can input a pathname
2047 of a file containing host IPs/names as the parameter value for the \-\-client option.
2048 For example, here is an example "host.list" file containing 2 hostnames:
2050 host1.your.dns.domain
2052 host2.your.dns.domain
2054 The fio command would then be:
2056 fio \-\-client=host.list <job file>
2058 In this mode, you cannot input server-specific parameters or job files, and all
2059 servers receive the same job file.
2061 In order to enable fio \-\-client runs utilizing a shared filesystem from multiple hosts,
2062 fio \-\-client now prepends the IP address of the server to the filename. For example,
2063 if fio is using directory /mnt/nfs/fio and is writing filename fileio.tmp,
2064 with a \-\-client hostfile
2065 containing two hostnames h1 and h2 with IP addresses 192.168.10.120 and 192.168.10.121, then
2066 fio will create two files:
2068 /mnt/nfs/fio/192.168.10.120.fileio.tmp
2070 /mnt/nfs/fio/192.168.10.121.fileio.tmp
2075 was written by Jens Axboe <jens.axboe@oracle.com>,
2076 now Jens Axboe <axboe@fb.com>.
2078 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
2079 on documentation by Jens Axboe.
2080 .SH "REPORTING BUGS"
2081 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
2084 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
2086 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob